BY 24 1957

Penicillin, protoplasts and bacterial L-forms
Joshua Lederberg

In the text-beok classifications of the living world, the Bavteria are
traditionally represented ae mimite organiams which reproduce exclusively
by transverse fission, This conception is engrained in the very classname
of Schiso-mycetes, or fission-fungi. We now know that these organisms are
more complex than our classification of them, For example, at least some
species can participate in a process of sexnal reproduction. Furthermore ,
even in asexual propagation, some bacteria under certain circumstances can
assume very unebacterial shapes and modes of growth, the L-forms which are
the subject of this article,

"L' stands for the Lister Institute for Preventive Medicine, at Lomon,
where in the hands of Dr, Emma KijenebergersNobel the more recent chapters
of our narrative began about 20 years ago, Many previous authors have des~
cribed bisarre, atypcial forms of bacterial growth, For the most part, the
earlier findings were ignored, or discounted, or rel@gated to footnotes in
the textbooks, (It sometimes happens in science, when upsetting new knowledge .
intrudes on comfortable habits of thought, that the first reaction is "pro=
bably untrue", the second "well, not very important" and finally, "well, so-
andeso discovered it years ago".)

Dr, Klieneberger was studying the life history of a bacterium called
Streptobacilius meniliformis which is involved in a disease called Haverhill
Fever, As the Latin name implies, this grows typically in the form of long
chains of bacilli (rods) which remain attached for some time after they have
increased in mmber by transverse fission, However, in some culture media,
~2e

she noticed a quite different growth form, which she designated *L," for
this instance, and Los L, etc. for similar forms in other cultures, Unlike
the typical bacilli, the L-form consisted of nondescript masses of gramles
and globules of all sizes, often distorted in shape, She made a concerted
effort to separate the Leform from the bacillary growth, and succeeded to
the extent of isolating true-breeding cultures of the Leforms, However, she
did not sueceed in isolating pure-breeding cultures of the bacilli free of
the L-forms, She suggested two alternative hypotheses, a) that the Leforms
were a stage in the life-cycle of the bacteria, and b) that they represented
another organism, symbiotie with the Streptobacillus, and not readily separated
from it, Another obscure group of organisms, the "pleuropneumia-like organisms",
"PPLO", was a likely candidate for the affinities of such a symbiont, on
account of certain morphological similarities, Her choice at the time was
the symbiont hypothesis (b), partly because of her partial success in separate
ing the bacillary from the Leforms, partly because this would put the least
strain on traditional views of the life-history and morphology of an otherwise
seemingly typical bacteriun,

During the succeeding decade, her stadies were complemented by those
of Dr, L. Diemes, at the Massachusetts General Hospital in Boston, Leforms
were isolated from a number of different bacteria, and in a few instances, he
was able to show by direct microscopy the transformation of bacilli into L~
forms and vice versa, Dr. Klieneberger-Nebel concurring in these observations,
there was no longer any doubt of the life-cycle hypothesis, that bacteria were
liable to another form of shape and growth, However, the biological signi-
fiecance of L-forms remained quite obscures they have been assigned such
functions as rejuvenation, adaptation to harmful enviroments, ani sexuality
(the last speculation being the initial source of my own interest in them).
A detailed analysis was held up for some time by the difficulty of reproducing
Leform growth and quantitating it, in common bacteria, particularly in species
like Egcherichia coli which are best vorked-over as materials of genetic
experimentation,

One of the reasons for thinking that L-forms might be adaptations to
noxious envirorments was the finding that they were quite resistant to peni-
cillin, though the bacillary forms from which they were derived and to which
they might revert were quite sensitive to this antitdotic. Penicillin hae
therefore been used syntematically both to eliminate bacillary forms and to
promote their transformation into the L-type of growth. By means of penicillin,
Dienea and many other workers produced Legrowth in a wide variety of organisms,
though not with the genetically standardised strains of E. coli, Three aspects
of the action of penicillin could be distinguished, at least in retrospect:

1) many bacteria were converted into fragile, globular bodies; in the same
cultures the majority of cells would customarily be killed and dissolve, 2)
In agar medium, containing penicillin the globular bedies could sometines
proliferate to give L-colenies like those pictured (figure ). 3) While
the Lecolenies would ordinarily revert to the bacillary form in the absence
of penicillin, occasionally a stable L-form could be isolated which would re~
tain ite peculiarities on standardised mediua, the same medium that would
support bacillary growth of the parent culture, Any comprehensive theory of
Leforns must therefore account both for penicillin-induced and genetically
stabilised L-type growth, We mst leave this thread of our narrative, which
now dates about 1953, with the anticipatory comment that most of the clues
meeded for the explicit statement of such a theory had already been collected,
though they were scattered through a large number of observations on many
species. It is equally remarkable that while the problem of the mechanism of
le

action of penicillin had been vigorously, but unconvincingly, attacked from
a biochemical point of view, the morphological effects of penicillin were
either ignored or put down to non-specific reactions to injury.

The impact of Sir Alexander Fleming, the renowned discoverer of peni-
cillin, hangs over this story, for (in 1922) he had also discovered lysozyne,
an ensyme found in egg-white, and secretions such as salive and tears, The
ensyme dramatically lyses certain bacteria, evidently by attacking their
cell walls, This conception of its action was extended by Weibull and by
Tomscik in 1953, with studies on another becteriup, Bacillus megeterium,

a species whose very large cells (a property acain coded in the name) make
it apt for cytological studies, Whereas lysosyme rapidly dissolves these
cells in ordinary media, the addition of large amovnts of sucrose, to give a.
concentrated medium of high osmotic pressure, protects the cells from complete
lysis, Instead, the dis-solution of the cell walls releases globular prete-
plasts, which are the soft living unit contained within the rigid cage. The
term protoplast is borrowed directly from botanical usage, where it means
precisely the same thing. However, unlike the protoplasts of plant cells,
bacterial protoplasts can survive only in concentrated media, Dilution inte
ordinary media constitutes an ‘osmotic shock’ which causes the protoplasts
to burst. The very ease with which the protoplasts can be broken has made
them excellent sources of intracellular constituents for biochemical studies,
and this alone has motivated studies on the development of similar procedures
for other bacteria, e.g. again E. coli.

The spherical form and the fragility of these protoplasts at once suggests
a possible relationship with Leforms. This was at first obscare, as the lyso-
syme~-imiuced protoplasts of 5B, megaterium have not yet been made te proliferate
into L~forms,
The lytic action of penicillin on growing bacteria had been used in
this laboratory for several years for the isolation of growth-factor- requiring
mutants, Since nonegrowing bacteria are left intact (an earlier empirical
observation), penicillin is used to kill the preponterance of cells that can
grow in a defined medium, ani leave behind those that are starved for specific
components in which the medium is lacking.

Since sucrose would forfend the lysis of Bacillus megaterium by lysozyme,
would it have the same effect on growing E, coli, ani thereby furnish a method
for producing protoplasts of this species?

The experimental answer to this question was indeed yes, In the presence
of penicillin, growing cells, inetead cf dividing by fission, enlarged and in
two hours or so popped out of their skins to release rather large (that is to
say, about 5 » or 1/5000ths of an inch across) spherical bodies, On account
of their shape, and their easily verified fragility to osmetic shock, these
are also considered protoplasts, It would, however, be difficult to say that
either these protoplasts or those obtained by lysosyme lacked every element of
the normal cell walls, a question that meds further study,

Bat are net these protoplasts in fact the same elements as have been
described earlier, as the initial stages of the l-form cycle? The chief novelty
in these experiments was the viewpoint that motivated them, and the fact that
the use of a protective medium allowed the transformation from rod to glole
to be quantitatively described. In addition, FE. coli had not previously been
amenable to this shift, The very reason that this species displays sensitivity
to osmotic shock in the most dramatic form, has also helped clarify the signi~
ficance of protoplasts,

The facte are now ready for a systematic interpretation, some of the
conjectural elements of which have since been confirmed, The mechanism of
action of penicillin is to inhibit the synthesis of cell] wall substance,
Unbalanced growth, namely of protoplasmic mass versus the wall, leads to

the rupture of the cells, The naked protoplasts will promptly lyse in ordi-
mary media, but can be protected in concentrated media, The fate of these
protoplasts depends on their enviroment, If penicillin is removed, cell
wall will be resynthesiszed and the normal shape and osmotic tenacity will be
restored, However, if the protoplasts are subjected to osmotic shock, or
allowed to grow too long in the presence of penicillin, they will turst, and
of course lose their viability, This contrast shows that no other action of
penicillin has to be invoked to explein ite lethal action, The unique
chemical makeup of bacterial o#ll walls accounts for the chemotherapeutic
specificity of penicillins the animal body lacks any targets which are
chemically similar to these walls, This hypothesis is not yet completes the
secom andi third aspects of penicillin L-form relationships will be considered
shortly,

This view of the biochemical target of penicillin as the synthesis of the
bacterial wall was not only compatible with existing data on its influence on
cellular permeability, but was soon corroborated by a completely different
line of evidence,

At the Department of Biochemistry (which is next door to Genetics at
Wisconsin) T, 5, Park and his professor, Me Johnson, had found that staphyle-
cocel exposed to penicillin accumlated a series of compounds related to
mcleic acids, but of a formerly undescribed structure, Later work by Park
established that these compounis were conjugates of Uridine-Diphosphate coupled
with amino compounds: an amino sugar (muramic acid) and certein amino acids,
The significance of these compounds was quite obscure at the time, and they
were related to the synthesis of moleic acids, which for various other reasons,
not now convincing, were thought to be the targets of penicillin action,
~ T=

Meamhile, other derivatives of Uridine diphosphate have been found to be
important, not so mach in nucleic acid metaboliom, as in the metabolism and
coupling of various sugars, That is, it is now more significant to regard
these compounds as activated forms of aminoesugars, rather than mcleic acid
derivatives. (It is quite possible that both pointe of view will be justified,
when we learn more of the role of meleic acids in controlling the synthesis
of proteins and other specific macromolecules.) Similtaneously with the
morphological studies just reviewed, Park and Strominger found that the
amino-residues were equivalent to those found in the cell walls of untreated
staphylococci, They could therefore propose, quite independently, that
penicillin blocked the further metaboliem of these activated anino-residues,
and furthermore that their normal metabolic fate is incorporation into wall
substance, However until the ensyme systems which are responsible have been
isolated, it is impossible to judge whether the primary inhibition by penicil-
lin is of a polymerising ensyme, on the synthesis of other building blocks,
or even on some more organized structure which is involved in wall~building,

What of the further growth and viability of the protoplasts? In liquid
medium, they grew rapidly in the sense that proteins, including new enzymes,
and meleic acids are synthesised, (In fact, they have been used, together
with partially disrupted fragments of protoplasts, for exciting studies on
the biosynthesis of these compounds by Dr. 5, Spiegelman at the University of
Tilincis), But when followed microscopically, the protoplasts actually do not
divide in liquid cultures. They enlarge enormously ani eventually burst, even
in a protective medium. However, most students of L-forms have emphasised the
role of agar in supporting this type of growth.

Initial attempts to grow the protoplasts of E, coli in penieillin agar
gave erratic results, which were finally traced to still mysterious strain
~Be

specificities, to the concentration of penicillin used, and to the enrichment
of the medium. Some cultures have simply given no L-type growth under any
conditions tried, although they are indistinguishable from the others in
forming the initial protoplasts, It was also found that levels of penicillin
which were quite sufficient to inhibit becillary growth still were not encugh
to support optimal Letype growth, Since there is then a sone of intermediate
penicillin concentration that permits neither, this will be referred to as

the sone effect. Fimlly, the addition of enrichments such as meat extract,

in combination with sucrose, gave quite high ‘efficiencies of plating' so that,
@.g., half the celis plated into an optimal medium might give L-colonies, as
might be counted when they developed fully. These colonies have a transparency,
slow growth and texture that distinguishes them from ordinary bacillary colonies
even at low magnifications, When they are studied at high magnifications,

they give the bizarre appearances shown in figure .

These colonies could be passaged in subculture by cutting out the blocks
of agar in which they were embedded, grinding them cautiously, and replating
the dispersion into fresh plates containing the same penicillin agar medium,
However, despite the thousands of globular elements that can be seen in such
an Letype colony, the yield of subculture colonies will be only from ten to
fifty; most of the visible elements are therefore inviable or extremely fragile,
Even after twenty passages under optimal cordlitions, all viable elements
promptly revertel to typical bacteria in the absence of penicillin, These
experiments had therefore not yet reached the third stage: genetic fixation,

The described effcets could all be referred to the inhibition of wall synthesis
in the uninterrupted presence of penicillin, and ite resumption when the
penicillin was withheld,
The manner in which protoplasts proliferate requires more study by re-
fined techniques, since the elements are so small as to be at the useful
limit of magnification of the phase contrast microscope, For this reason,
the picture to be described may be incomplete, It appears, however, that
the protoplast is a semifluid droplet, which changes ite shape in response
to ite own surface tension, In agar, however, it will be compressed by the
matrix in which it is embedded, and especially as it enlarges, it extenis
into the spaces in this matrix, These extensions or blebs, may remain minute
gramiles, become long or fingerlike, or may round up immediately, but round
up they will until they actually pinch off from the mother protoplast,

Some of the daughter forms, perhaps thoee which have received a micleus

along with a blob of protoplasm, will contime to grow and generate their

own progeny, At a glass surface, a developing L-colony has been seen to
spread along the glass and break up into droplets of varying sise, exactly as
expected on this picture, The semi-fluid protoplast thus has no regular
mechanism of division, which normally depends on the wall, but is at the mercy
of external pressure and surface tension which can cause it to fragment as

it grows in a confining milieu, As another aspect of L-type growth some
authors have stressed the viability of minute granules, some perhaps half the
sise of the normal bacterium. It has been very difficult to get direct evi-~
dence of the viability of such a particle, which would be nearly subaicroscepic;
indirect measurements besed on the occasional passage of viable elements
through filters of graded size may be influenced by the plasticity of the
protoplasts, which might enable them to be squeezed through pores that would
restrain a more rigid particle, However, the possible 'filtrability! of Le
form granules is more a technical question than anything else, so long as the
limiting dimensions wtill leave room for packaging the minimum apparatus for
genetic contimity of the bacterial cell.
10~

Perhaps the most puzzling feature of L-forms, is the third aspect, the
genetic fixation of L-type growth after prolonged passage in the presence
of penicillin, The leading clues for this came from still another source,

By and large, the chemical makeup of all cells is startlingly uniform.
The same array of amino acids is basie to the structure of proteins in bacteriz
as inman, But there are exceptions, amd one of the clearest of these was
the discovery by Dr, Elisabeth Work, in London, of a new amino acid, DAP or
diwaminopimelic acid, which has been proved to occur in a variety of bacteria,
and nowhere else in nature, Further studies have localised DAP entirely or
nearly entirely in the cell walls of the various bacteria, Dr, B, D, Davis
of New York isolated a mitant of FE, coli which required DAP for growth, and
this mutant has been very useful both as a biological assay for DAP and for
the further elucidation of its function, In a striking corroboration of the
hypothesis for penieillin action, when cells of the DAP-leas mtant are
deprived of DAP in an otherwise rich medium, they undergo the same transfor-
mation tnto protoplasts as the wild type strain does in the presence of peni-
eillin, The DAP+less proteplasts are osmotically fragile, ani will revert
to normal bacillary forms if they are furnished with DAP, It ia therefore
inferred that unbalanced growth, in which wall-synthesis is now impeded by
the lack of this essential building block, while other aspects of protoplasnic
synthesis contimue, All this is exactly analogous to the effect of penicillin,
which however affects another aspect of wall formation,

In addition, when DAP-less cells are plated into enriched agar, whether
penicillin is present or not, they proceed through the same cycle of L-type
growth, Since ordinary media are deficient in DAP, this mutant would be
recognised as a stabilised L-form by the criteria applied to those isolated
by other workers, We therefore are led to the conclusion that stabilised
wh Le

Leforms are mtant strains which have suffered a genetic block to wall for.
mation, This bleck need not be in DAP-metaboliam, nor at the still unknown
target of penicillin action, Any mitational block to an essential part of
wall synthesis should give the same result, In fact, the stabilized L-forne
isolated by other workers have not responded to DAP, and must therefore have
another defect, However, the defect would have to be for a function unique
to wall formation, since one that influenced overall growth would inhibit
pacillary or Letype growth alike,

There remains the question of why such wall-defect mutants should tend
to cccur in the presence of penicillin, Is penicillin iteelf mutagenic, or
does the cell-wall contain or depend on a self-perpetuating structure? These
notions are not readily ruled out, and in certain contexts they might be
plausible, but an even simpler explanation is possible,

The sone effect of penicillin was mentioned previously. That is, certain
levels of penicillin discourage both baciliary and Legrowth, In fact, even
to very high levels, penicillin continues to s timlate Letype growth. On
the other hand, the DAP-less mutant produces good Letype growth at any level
of penicillin, This suggests that L-type growth is best when wall formation
is completely inhibited, either by saturation with penicillin or by the super
imposition of a genetic bleck, Therefore during the passage of L-type colonies
in the presence of penicillin, any random mutation imposing a genetic defect
on walleformation, will have a selective advantage and will tend to overgrow
its neighbors, Since quite large populations of cells are always involved in
these experiments, even mutations occurring at quite ordinary, that is low,
rates, can lead to the results found, It is quite possible, however, that
some bacterial species have much higher rates of mtetion than others for the

occurrence of such defecte,
~L2~

By the application of these principles, it has been possible to isolate
new, ‘stable! L-types from E, coliy these have so far proved to be mutants
requiring DAP, but the same rationale is expected to lead to the detection
of other metabolic defects, e.¢, for the amino-sugers mentioned previously
which scem also to be unique to bacterial walls,

What then of the functioml significance of the bisarre L-type or, as
may now be more appropriate to call it simply, protoplastic growth? This has
been tested by genetic means, and found to be irrelevant to sexual reproduction
(which other experiments have related to the pairwise conjugation of intact,
fall-walled bacteria). It now appears to be nothing more than a sign of the
nakedness imposed either by external inhibition or internal defect. However,
the bacterial wall has many functions, of concern both to the microbe and the
microbiologist, For example, it is the site of attachment of certain viruses,
so that the protoplast is essentially resistant to their action, As these
viruses also carry a wall-dissolving ensyme similar to lysosyme, they may
play some part in the occurrence of protoplastic forms in nature, But at this
instent, the naked forme sre parhaps of most interest as laboratory tools,
both to study the function of walla, and for hopeful attempts to elude them.
Insofar as the wall may be a barrier to the movement of substances we may wish
to import ami export, ites removal from a still viable unit is a hopefal
stategem in the scientific siege of the Cell,

In this account, I have indulged in name-dropping to good purpose, It
is perhape more obvious, but no truer of this subject that its development has
depended on the discoveries and insights of unnumbered researchers working
throughout the glebe. It is no accident that a Swede working in California,
an Austrian in Massachusetts, a German in Lorion and a Hungarian in Switserland
are among the leading characters in this memoir,
References for further information
[Thin list is selected from works in English only]

Dienes, Le and H. J. Weinberger 195k ‘The L forms of bacteria.
Bacteriological Reviews. 15:245-288. Dec

Work, B. 1957 Biochenistry of the bacterial cell wall.
Nature 179: 841-847. Apr 27
Symposium on Bacterial Anatomy. Society of General Microbiology, Gt. Brit.
Cambridge Univ. Press, 1956

Park, J.T. and J. L. Strominger 1957 Mode of action of penicillin.
Soience, 125: 99-101. Jan 18

Lederberg, J. and J. St. Clair 195%

Journal of Bacterlology